perimeter — Architectural Boundary & Bypass Detection Engine

Domain Rubric

1. Boundary Audit Dimensions

• Sanctioned Primitives vs. Raw Bypass (Horizontal): Identify internal wrappers governing I/O, network clients, resources, or lifecycles; flag call sites bypassing them for raw runtime APIs.
• Layer & Dependency Directionality (Vertical): Trace import graphs across architectural layers (Domain, App, Infra, UI); flag inverted or cross-boundary imports.
• Resource & Sandbox Hygiene: Audit storage paths, concurrency pools, and memory buffers; flag unanchored filesystem access or unthrottled resources.

2. The Load-Bearing Qualification Gate

1. Existing Sanctioned Primitive: Codebase defines an internal wrapper, base class, or canonical path addressing the boundary.
2. Empirical Bypass Evidence: At least one active call site violates the boundary by calling raw APIs directly.
3. Quantifiable Failure Risk: Bypass induces architectural drift, resource leaks, memory exhaustion, or security exposure.

• Negative Boundary: Forbid declaring constraints for cosmetic styling, trivial utilities (

  Math.*

  , collections), or speculative rules lacking active primitives.

Canonical Output Contract

# Perimeter Report: <Target Subsystem / Scope>

## 1. Executive Verdict
- **Verdict**: <HARDENED | PARTIAL | UNPROTECTED>
- **Scope Inspected**: <Directory, Subsystem, or Package>
- **Active Boundaries Detected**: <Count of qualified boundaries>
- **Summary**: <Single-sentence technical judgment on architectural boundary integrity>

## 2. Qualified Boundary Declarations
*(If HARDENED: 'All architectural boundaries are enforced by active test suites; zero raw bypasses detected.')*

### Boundary: <Constraint Name, e.g., Banned Direct Network Client>
- **Risk / Failure Mode**: <Concrete failure risk>
- **Offending Evidence**: [<file>#L<N>](file:///path/to/file#L<N>) ── <Bypass snippet or call site>
- **Codified Rule**:
  - Banned: `<Forbidden raw API, symbol, or import pattern>`
  - Required: `<Sanctioned internal primitive or approved route>`
  - Whitelist: `<Exempt wrapper files or 'None'>`
  - Recommended Tier: `<Script Guard | Arch Unit Test | AST Linter>`

## 3. Disqualified Invariants (Pruned)
- **[PRUNED] <Candidate Pattern>**: <Reason for rejection, e.g., No existing codebase primitive / Trivial style issue>

## 4. Pipeline Routing
- `/forge` ── Authorize and implement automated architecture tests for qualified boundaries.
- `/clarify` ── Boundary ownership, whitelist exceptions, or architectural trade-offs require human deliberation.

Execution Protocol

Phase 1: Macro Anchor Grounding

1. Ingest target directory or subsystem.
2. Scout high-level architectural anchors (internal wrappers, sanctioned clients, resolvers, base interfaces) using symbol and file search tools.
3. Establish grounded anchor context to seed downstream inspection.

Phase 2: Scale-Adaptive Boundary Audit

1. Evaluate target scope scale:
   • In-Turn Execution (1–5 files or localized module): Audit call sites against discovered anchors directly in-turn.
   • Subagent Fan-Out (broad codebase or cross-layer scopes): Launch 1–3

     research

     subagents (Role:

     Perimeter Reviewer (<Slice>)

     ). Seed each with:
     • Target slice directory.
     • Discovered architectural anchors from Phase 1.
     • The Load-Bearing Qualification Gate and path to

       references/fitness-functions.md

       .
     • Canonical sub-report contract requiring offending code pointers (

       file:///...#L<N>

       ).

Phase 3: Codification & Turn-Halt Gate (Fan-in)

1. Reconcile findings into the Canonical Output Contract, moving ungrounded rules to Section 3 (Pruned).
2. If boundary ownership or whitelist exemptions contain unresolved trade-offs: halt and deliberate (via

   /clarify

   ).
3. Emit the report into the conversation stream and halt turn immediately.